It is often difficult to decide whether a given form belongs to one or another of these two groups. I have tried to show that the best and strictest conception of varieties limits them to those forms that have probably originated by retrograde or degressive steps. Elementary species are assumed to have been produced in a progressive way, adding one new element to the store. Varieties differ from their species clearly in one point, and this is either a distinct loss, or the assumption of a character, which may be met with in other species and genera. _laevifolia_ is distinguished by the loss of the crinkling of the leaves, _brevistylis_ by the partial loss of the epigynous qualities of the flowers, and _nanella_ is a dwarf. These three new forms are therefore [565] considered to constitute only retrograde steps, and no advance. This conclusion has been fully justified by some crossing experiments with _brevistylis_, which wholly complies with Mendel's law, and in one instance with _nanella_, which behaves in the same manner, if crossed with _rubrinervis_. On the other hand, _gigas_ and _rubrinervis_, _oblonga_ and _albida_ obviously bear the characters of progressive elementary species. They are not differentiated from _lamarckiana_ by one or two main features. They diverge from it in nearly all organs, and in all in a definite though small degree. They may be recognized as soon as they have developed their first leaves and remain discernible throughout life. Their characters refer chiefly to the foliage, but no less to the stature, and even the seeds have peculiarities. There can be little doubt but that all the attributes of every new species are derived from one principal change. But why this should affect the foliage in one manner, the flowers in another and the fruits in a third direction, remains obscure. To gain ever so little an insight into the nature of these changes, we may best compare the differences of our evening-primroses with those between the two hundred elementary species of _Draba_ and other similar instances. In doing so we find the same main [566] feature, the minute differences in nearly all points. V. The same new species are produced in a large number of individuals.